GearGrabber is the AI receipt-capture flow that powers inventory + financial logging across every Crescender product. Take a photo of a receipt (or an invoice, or a delivery docket) and the AI extracts the structured data into your records in seconds. The same engine runs on the web, on the mobile app, and inside Creduca Asset for school-scale inventory operations.
It's the most quietly transformative feature we ship, turns the most painful music-business task (data entry from paper) into something you do without thinking.
What you point your camera at
Music shop receipts (printed and handwritten). Online order PDFs from Sweetwater, Thomann, Allans, Music Junkie, Better Music, etc. Delivery dockets when gear arrives separately from the invoice. Service receipts from repair shops. Lesson-fee receipts you give your students. Course-fee receipts for professional development. Anything with line items, prices, and a vendor.
Image quality doesn't need to be precious, a phone-camera snap with reasonable lighting is enough. The vision model handles glare, rotation, perspective skew, and handwritten dollar amounts surprisingly well.
What the AI extracts
Vendor name and contact details. Date of purchase. Currency (defaults to AUD for Australian users but multi-currency-aware). Each line item with description, quantity, unit price, and total. Tax (GST in Australia; other tax regimes for non-AU users). Order or invoice number. Any serial numbers visible on the receipt (often present for high-value music gear).
After the parse, you see the extracted data on a confirmation screen with the original receipt photo alongside. Most fields auto-populate correctly; fields the AI was uncertain about are highlighted for review. One tap to accept.
What happens automatically after you accept
Three things, in parallel. The financial side creates a transaction with the correct vendor, date, total, GST-handled, attached to the original photo for audit purposes. The inventory side adds each line item as a new gear record with photo, vendor link, purchase date, and serial number (if extracted). The events side detects ticket purchases (concert tickets, festival passes, masterclass admissions, exam fees) and creates calendar entries with the relevant dates.
All three sides are cross-linked. From the gear record you can jump to the source transaction; from the transaction you can jump to the gear it bought; from the calendar event for a course you can jump to the receipt that paid for it. Everything is one click away.
How accuracy is maintained
The AI is good but not perfect. The confirmation step is mandatory, you always see what the AI extracted before it commits to your records. Corrections you make feed back into the model's prompt context (your prior corrections are remembered, so the AI gets better at your specific vendor mix over time).
For high-value items (serial-numbered instruments, large invoices), GearGrabber prompts you to verify the serial number against the physical item before accepting. The few extra seconds catch the rare-but-painful case where the AI mis-reads a serial.
